Analog computing might get a new lease on life thanks to metamaterials that control electromagnetic waves. Researchers from the University of Pennsylvania (UPenn) developed a metamaterial whose tiny structures manipulate light to solve mathematical equations.

“Photonic calculus” encodes parameters into the properties of an electromagnetic wave. Once the wave passes through the metamaterial, its properties get altered by nanostructures.

The modified electromagnetic wave serves as the output of the light-based computing process. It carries the solution to an integral equation.

The UPenn researchers recently presented their metamaterial device by conducting a proof-of-concept experiment. They took advantage of the long wavelength of microwaves to build their macro-scale prototype.

The researchers said that the same principles behind their findings can be applied to shorter light waves. This means that their device can be scaled down to the size of a microchip.

Refined versions of the metamaterial device may serve as analog computers that run on light instead of electricity. These devices will be able to answer integral equations much faster than digital computers while using less energy. (Related: Powerful lithium-air batteries may soon “take batteries to the next level,” reveals study.)

A modern, nanotechnology-enabled take on analog computing

UPenn researcher Nader Engheta and his colleagues sought to modernize analog computing. In 2014, they released a tentative proposal for photonic calculus in the journal Science. In it, they demonstrated how a metamaterial could execute math operations on the properties of a wave that passes through it. Their latest physical experiments confirmed their earlier theory and introduced practical applications for metamaterial-based analog computers.

Engheta and his team built their metamaterial device out of polystyrene plastic. They used a CNC milling machine to create a network of air holes across the dielectric material. The holey appearance earned the material the nickname “Swiss cheese.”

The precise pattern of hollow areas in the dielectric material is pre-determined to solve an integral equation with a given kernel. Kernel refers to the part of a math equation that describes the association between two variables.

The phases and magnitudes of an electromagnetic wave serve as arbitrary inputs for the equation. When a wave enters the dielectric device, it gets inputted into the predetermined equation.

For example, a user may get tasked to map out the acoustics of a concert hall or a similar space. To answer this problem, he would need to create an integral equation and enter the data to get the answer.

The user can create the problem-solving kernel by milling a pattern in the dielectric metamaterial block. The sources of the sounds, as well as the other features of the room, will serve as the inputs, represented by the equation’s kernel.

The user can encode these parameters into the electromagnetic wave and send it into the kernel. The emerging wave will have the solution.

New metamaterial computing device works faster and uses less power than digital computers

The UPenn metamaterial device occupies an area of two square feet. Using the length of a microwave as the unit of measure, it spans eight wavelengths by four wavelengths.

“Even at this proof-of-concept stage, our device is extremely fast compared to electronics,” Engheta explained. “With microwaves, our analysis has shown that a solution can be obtained in hundreds of nanoseconds, and once we take it to optics, the speed would be in picoseconds.”

He and his team plan to shrink the metamaterial device so that future versions may fit on a microchip. They will also switch their choice of electromagnetic waves from microwaves to shorter light waves.

A photon-based, micro-sized metamaterial device will be able to match the performance of multi-purpose digital computers that supersede analog technology.

The researchers believe that they can use rewritable CD technology to make new “Swiss cheese” patterns on metamaterials. This will allow people to 3D-print their own custom-built analog computers.

Sources include:

Nanowerk.com

Science.ScienceMag.org